In recent years, considerable progress has been made in the development of computer programs that perform an analysis of digital images of human faces in order to identify the individual or individuals depicted in the image. However, varying conditions that may exist at the time the digital image was taken can interfere with an accurate performance of a face recognition task. For example, varying illumination conditions existing at the time an image is recorded can result in shadows cast across a face depicted in the image. Such shadows can prevent a computer system from performing an accurate analysis of facial features depicted in the image, and thus, cause a failure to properly identify the individual. Accordingly, there is a persistent need for the development of accurate and efficient techniques that can be utilized in the operation of computers relating to images, to provide improved performance in, for example, a face recognition method.